Christopher Baer (@christopherbaer) header
Christopher Baer (@christopherbaer) profile picture

Technical product in Silicon Valley. Looking for telos in the era of infinite capability. (Posts hand-written) Ex Wish, TripAdvisor, eBay, Stanford.

SF Bay Area, CA
Joined May 2008
Photo by Christopher Baer (@christopherbaer)Photo by Christopher Baer (@christopherbaer)Photo by Christopher Baer (@christopherbaer)Photo by Christopher Baer (@christopherbaer)Photo by Christopher Baer (@christopherbaer)Photo by Christopher Baer (@christopherbaer)Photo by Christopher Baer (@christopherbaer)Photo by Christopher Baer (@christopherbaer)Photo by Christopher Baer (@christopherbaer)Photo by Christopher Baer (@christopherbaer)
Remarkable email for us to see.
Sacha Baron Cohen emails Sheryl Sandberg November 16, 2016
17
This is a very astute summary (from the engineer who coined “open source” I believe) of the central issues in clinical medicine and public health right now: how to ration care, how to think about health equity/justice, what is socially vs individually determined.
So I recently posted on here that one of the reasons I'm opposed to an AI pause is that I don't want to die. An entire thundering horde of blithering idiots immediately started screaming at me that I'm a selfish bastard who wants to burn the world down. These people seem to assume I have some kind of supervillain bunker I plan to retreat into, cackling madly and clutching my extended lifespan to my chest, rather than having to live in the same future world as everybody else. This is a deranged response. I had to think for a while to even begin to mentally model the kind of bugfuck craziness that could produce such thinking. Now that I think I have done so, I have a public service announcement to make: Some people will try to convince you that life extension, or wealth in general, is a zero-sum game. Then in order for some to win, others must lose. For any individual person to achieve an extremely long lifetime, the world must be plunged into misery for others. These people are not your friends. They are not saying these things to help you or save you. They are telling lies which, if you believe them, will control you. And control is their goal. Did we learn nothing from the last century? The self-anointed scream that a vast catastrophe is heading towards us all, unless we surrender our wealth and liberty to them. The label on the catastrophe changes: overpopulation, pollution, global warming, rogue AIs. The prescription never changes. It's always a grift, always a lie, always a power-grab. Don't get fooled again.
15
Just downloaded this 499 page book on robophilosophy and it has a gripping start: "Whether we acknowledge it or not, we are, in fact, in the midst of a robot invasion. The machines are everywhere and doing everything. They may have begun by displacing workers on the factory floor, but they now actively participate in many aspects of intellectual, social, and cultural life." !!
"Robophilosophy: Philosophy of, for, and by Social Robots" @mitpress is now out and available. This book has been ten years in the making and includes ground-breaking research from the @Robophilosophy_ Conferences. Get it open access at: direct.mit.edu/books/oa-edit…
1
2
106
Very neat. Strange, but neat.
in case you were curious: fable 5.2 and I confirmed that if you record yourself singing and whistling every note in your range, you can play basically any song using only your voice no song is playing in this video. every sound is one of 64 clips of me holding a single note
1
24
CI/CD efficiency is definitely one of the things that is always worth checking on, often weekly, for vibe coders like myself. I was up to 10k minutes/month in GitHub Actions usage at one point, since LLMs tend to not optimize much here unless directly asked to do so. In this case, Dan reduced wall-clock duration by 65% with a one line fix: “ •apps/e2e/playwright.config.ts: run 4 Playwright workers on CI instead of 1 (local runs keep Playwright's default)…” What a disaster. I feel like GitHub should be helping users with this type of stuff. The entire “self-hosted” GHA ecosystem shouldn’t need to exist except for big biz customers. 😵‍💫
only one person (@foundmalek) replied with their OSS repo, so I took a look: a one line change reduced their CI time by 65%! PR'd and merged 🚀 you can just do things
37
Now that’s a good story.
Replying to @mdonnino
Dr. Raymond Adams (considered father of neurology) & colleagues did large case series putting forth link between thiamine & WE in 1950/60s though clinical trials were less rigorous at the time. Shortly before he died, I met him (age 96) & he shared a bit of a bombshell with me 2/
27
Thanks for explaining, and stepping back for a second, this is shockingly bad — nobody should ever use bitly short URLs if they are apparently long term compromises. What was bitly thinking?!
Ok time to ruin it for everyone by telling people how this work. Old bitly links no longer work, just go back and look for a big account post with a dead bitly link. Bitly allow custom urls if it's not taken already. You register it and now the post links to your stuff and you get free internet points.
1
1
121
Update
Earlier today, we identified an issue that allowed bad actors to repurpose certain old, inactive links to redirect people to content unrelated to the original posts. We have suspended the accounts involved and are disabling the affected links. Our Trust and Safety team is working to address the issue and strengthen protections against similar misuse in the future.
Readers added context they thought people might want to know
Bitly documentation states that bit.ly back-halves cannot be re-used even after account deletion, yet a similar vulnerability allowed repurposing of old links as documented in a February 2025 case Bitly claimed to have fixed. support.bitly.com/hc/en-us/artic… disruptionist.com/p/crypto-scamm… ransomnews.com/burger-king-tw…
19
Harvard prof with a very good review of all the different terms and concepts related to aging… precision of definition here helps a lot IMHO.
Our new paper, “Simply Aging,” is out in @PNASNews! What is aging, when does it begin, how do we measure it, and what would it take to extend lifespan? We tackle these questions directly and simply. 📄 Read: pnas.org/doi/10.1073/pnas.26… 💬 Discuss: simplyaging.xyz/
19
Quite a result; we really still don't know basic facts about what the brain does (in this case, mouse brain functional neuroanatomy).
I still can’t believe it. My PhD project of 8+ years at @HillmanLab, @ZuckermanBrain and @ColumbiaBME was published today in @NatureComms! We present a newly discovered body map in the secondary motor cortex (M2) of the mouse brain.
25
Cool new preprint detailing Univ of Michigan brain model they trained on 150 TB of brain scans, if I understood correctly arxiv.org/html/2611.15094v1 Open weights model I believe!
A new brain imaging benchmark was just released on @huggingface 🩻 🧠 NeuroBench is a 2.74 TB dataset of MRI and CT studies to test how well vision-language models actually read brain scans. This kind of benchmark is so important as medical imaging data is often private. With that public, anyone can check how models really perform. huge kudos to @mlins_lab for this release 🚀 🤗 dataset: hf.co/datasets/mlinslab/neur… 📜 paper: hf.co/papers/2611.15094
1
1
48
Might be worth trying to do some brain surgery on the brain model:
Want to try weight-space programming? Today we're releasing dspy-wsp, a prototype weight-space optimizer that works like any other @DSPyOSS optimizer. Using dspy-wsp, and about ~$5-10 of Astra, we improved every decision model we tested. cmpnd.ai/blog/weight-space-p…
1
24
This is like openclaw’s pete when he was spending over $1M/month in inference — way overkill, but paves the way for the rest of us to have good economical solutions later. Decentralized, resilient utilities (electricity, compute) make a lot of sense to me. Stronger overall grid, deeper integration.
I’m going a step further. I already have 12kW of roof solar but I’m about to add 3kW more and batteries to my mini-datacenter (shed). My compute will be in a cooled & walled off space. 2x10G Fiber from the house to the shed. Here is the roof pre-install. Will update over the course of the project.
1
52
Is this true? Dartmouth halting math PhD admissions? If so, that’s a tragedy. This is such an exciting time for knowledge and discovery…
119
Yeah. A lot of politically-mandated technical things may end up being self-defeating as algorithms / data science continues to progress. That isn’t to say don’t do things, but simply that there are tradeoffs here and it’s important to be clear-eyed that these are difficult problems, with inconsistent solutions. If, from a policy perspective, that is fine — well then, okay.
The European Commission is ordering Google to share the search data of all Europeans with other companies (AI chatbot makers, search engines, etc.), after a kind of "anonymisation". I tested it on the real Google data, hundreds of billions of search records. I broke it. Findings. I singled out the people behind half a million searches, and that only for few countries over one month of August (record isolation). I joined the pieces of search history beyond the defined cap of 3, more than 200,000 sets just for Germany and only in August (linkage). I also confirm that it is possible to perform inference on the dataset, for example concerning health, religion or political opinions. This moves us beyond data protection, and in the national security direction. My report: "Alone among a thousand". (link below)
1
1
36
Disclosure: I only read the tweets, not the underlying report. Seems consistent with prior concerns that specific individuals can be identified on Reddit even if anonymous there, because LLMs can profile based on other public writing.
10
I am entertained because this makes no sense at all, and yet in practice is totally fine… because at this point I only use the cf CLI for all Cloudflare config (via LLM). So the actual human-facing Cloudflare Dashboard is no longer important, and can have whatever wacky experience 😂
What if you could explore Cloudflare's products as a universe? cloudflare-universe.jilles.f…
1
163
Very good point. If you are building a software factory, make sure to use best practices for factories: continuous improvement, just-in-time manufacturing, etc. (See comment on original post for the other standards.)
If you haven’t, please read about what Toyota did in manufacturing the mid 20th century. they solved a bunch of the problems people are having managing agent swarms. And they wrote it down as generalisable principles
1
34
Interesting how there can be security risks at every step of LLMs: * webpages (for sure), PDFs and images could (in theory) have hidden prompts in them for LLMs * untrusted open-weight LLMs could have hidden actions that are triggered with specific keywords etc * inference providers that claim ZDR may still be saving and selling data and/or be hacked We’ve been really lucky for the last couple decades on how safe Mac ecosystem is (in general), but I think we are back to dot com era of needing anti-virus type software that is actively trying to find these new classes of problems. Grok bot has been notifying me it is ignoring hidden instructions for LLMs, so there is a certain value in paying the subscription fees…
Par souci de sécurité, on ne dépose jamais une image récupérée ainsi sur un chat LLM. L'image peut contenir des instructions invisibles pour un oeil humain, mais parfaitement lisibles pour le LLM. C'est du même genre que scanner un QR Code trouvé dans la rue.
1
58
Same
This is extremely relevant and the reason why I tend to use open weights in their original format and quantize by myself when needed. Check the source!
1
8